Here’s the second piece of the frame, now the structure of the scooter is basically complete. I can even run and jump on it and roll for a few feet before falling off. It’s like riding a hand truck. I did the mig-welding myself today!
I finished the first part of the frame, which consists of forks to hold the wheel axle, and a mount for the motor. The whole thing is going to be made out of 0.75″ square steel tube, since this is the easiest material to cut and weld at various angles. Round chromoly tubing (like what BMX frames are made of) would have been better because it is lighter and stronger, but it is too difficult to cut and weld joints. The whole frame is probably going to be really heavy when it’s done because the square steel is so bulky, but it’s so easy to work with that I did all the cutting and milling of it myself, and Don easily mig-welded the joints.
I still have to figure out where the handle bars are going to go, and how to mount the electronics and power supplies.
The english translations in China are by far my favorite part of traveling here.
-Punish Severely All Sorts of Crimes
- I 
Dancing
-California Beef Noodle King USA!
-Hard Speed Honey
-Kiss ‘N Bake
- Scalet Model
- Roast Duck Coffee
-You want beautiful Chinese Girl?
My favorite, though, is this new shirt - “Different Kicker - Soccer Hero”
The god that the people venerate does not carry a cross… ask the English… our god carries the 10 on the back.
I start my day off right, the Beijing way, with a cup of tofu-juice from a street vendor!
Now that I recognize all the bus and subway stops, traveling is easy - even mundane, since the feeling of being lost was a significant part of the excitement from the first day. I went to the conference to get a feel for the timing and expectations of quality from the presentations.
The posters at ICME ranged from the merely bland,
to the depressing mediocre that would blend in with the average middle school student science fair display,
to the comically immature.
Since it was raining most of the day, I tried to work on my presentation for Thursday. I ended up getting distracted and mostly just read the somethingawful forums and Digg all afternoon, and my presentation still isn’t finished.
We went back to the Turkish restaurant for dinner.
Bonus Riddle Time!
Q: Why did Yusuf and Andrew order a dinner for four people and an entire carafe of white wine?
A: Because Yusuf already ate a big lunch and doesn’t drink alcohol.
I tried to draw the Ottoman Coat of arms, but I think it looks better as a bird.
Summary: I woke up late, tried to do some work, read the internet instead, and spent too much money on a huge dinner at a restaurant. Might as well have been in Orlando…
Yusuf and I arrived in Beijing after a long but uneventful plane ride. We spent the next day walking around trying to find key locations such as our hotel, the convention center, a halal restaurant, and a universal power adapter.
Our hotel is in the vicinity of the Beijing Railyway Station. We had two maps for our hotel, but they gave conflicting locations. None of the street signs were in English letters, and none of the cab drivers and fruit/dumpling/dvd/magazine sales associates spoke any English. My laptop was dead, and my cell phone can’t dial out.
We walked past the train station,
through a back alley,
around the police station at the corner,
and found our hotel hidden away on the left.
The hotel room is roughly the size of my dorm room.
At night there is a view of the courtyard made eerie by the green glow of the lamps.
Early in the morning, everyone is out exercising - not just jogging, but practicing Tai Chi, walking backwards, gyrating at the hip, and raising one’s leg above the shoulder and slapping the thigh vigorously.
Due to a regrettable administrative oversight - there are actually TWO Crowne Plaza Hotels in Beijing, one across from the convention, and another in an entirely different part of the city - we had to find transportation to ICME. Unfortunately, no one on the street recognized the phrase “Beijing Convention Center”, and we had no street address. We bought a map, but it had no English letters. We drew a circle around the general area we thought we needed to go and let a taxi drive us 45 minutes through traffic to the North.
We couldn’t see the convention center or hotel from where the taxi let us off, so we just walked in circles for two or three hours. Behind a dirty alley, we found a construction site where some women were cooking food to sell to the workers. For just under a dollar, I bought myself an excellent lunch of stir fry vegetables and pork, seaweed, and almost a liter of beer.
Yusuf and I looked for the convention center by using a search strategy consisting of pointing at a tall building, suggesting “maybe that’s the conference hotel,” and walking a half hour in that general direction - which surprisingly paid off over time.
Today was just a rehearsal, though, because they just had special tutorials which we were not registered for. It was much easier to get back to the hotel by taking the bus and subway, since a nice lady at the conference wrote down the route names for us in Chinese characters. After a brief rest, we went out for another walk. We had dinner at a nice kosher Turkish restaurant, and I luckily found a power adapter.
The gyro rate sensor on the Sparkfun IMU has a broken gyro sensor. The voltage would go out on one or the other axis intermittently. I decided to switch to using the IMU that came with Don’s Rotomotion Rev2.4 controller (http://www.rotomotion.com/prd_REV2.4.CTLRA.html).
I wasn’t able to program the controller board itself. The chip on the controller is an ATMega32, but it wont respond to any attempts to load a program. The IMU itself works fine with the Robostix controller. Since this actually takes the 5 volts from the AVR, I had to build another connector with no need voltage divider.
My Macbook now crashes every time I change the serial port properties in Minicom, and I have to hard reboot. No idea why, but I’ll use Hyperterminal on my desktop.
I wrote a Kalman filter in MATLAB, and captured a bunch of data from Hyperterminal to play with. I was able to generate this snazzy plot, which illustrates the essence of a good filter. The accelerometer is very noisy, but the gyro is very precise. Unfortunately any error in the gyro gets accumulated hundreds of times a second, so it starts to drift. The ideal signal will cling to the accelerometer, but retain the shape of the gyro integration.
To make it easier to visualize the result, I made a program that reads in the filtered angle from the AVR and draws a little unicycle pendulum on the screen. Lag and gyro drift are a lot more apparent this way. It was a pain to read from the serial port in Windows. For some reason the regular Win32 calls to read from the com port crashed, possibly because of the USP-serial adapter I’m using. Since Cygwin can read from it easily, I had a Cygwin program and the Windows program communicate to eachother by constantly opening and closing a text file. A stupid technique to resort to, but it worked without any problems.
Everything on the Robostix runs at 5V. The Sparkfun IMU requires 3.3V. Don was too busy to help me find a voltage regulator, so I’m decided to make a voltage divider.
I cannibalized some 100ohm resistors from a toy logic latch I made on a breadboard a few years ago withtransistors and resistors from RadioShack.
I also smashed apart a Apple remote to see if it had a 3.3V regulator, but none of the surface mount parts with three prongs had enough letters to look them up.
I spliced two pieces of ribbon cable with a disugsting knot of resistors.
Under no load I get a sweet 3.3 volts. With the IMU connected, it drops to 2.7, which is below spec for the accelerometer. It seems to work, so I’ll leave this alone for now.
